mercurial/verify.py
author Anton Shestakov <av6@dwimlabs.net>
Fri, 07 Jan 2022 11:53:23 +0300
changeset 48687 f8f2ecdde4b5
parent 47877 2174f54aab18
child 48875 6000f5b25c9b
permissions -rw-r--r--
branchmap: skip obsolete revisions while computing heads It's time to make this part of core Mercurial obsolescence-aware. Not considering obsolete revisions when computing heads is clearly what Mercurial should do. But there are a couple of small issues: - Let's say tip of the repo is obsolete. There are two ways of finding tiprev for branchcache (both are in use): looking at input data for update() and looking at computed heads after update(). Previously, repo tip would be tiprev of the branchcache. With this patch, an obsolete revision can no longer be tiprev. And depending on what way we use for finding tiprev (input data vs computed heads) we'll get a different result. This is relevant when recomputing cache key from cache contents, and may lead to updating cache for obsolete revisions multiple times (not from scratch, because it still would be considered valid for a subset of revisions in the repo). - If all commits on a branch are obsolete, the branchcache will include that branch, but the list of heads will be empty (that's why there's now `if not heads` when recomputing tiprev/tipnode from cache contents). Having an entry for every branch is currently required for notify extension (and test-notify.t to pass), because notify doesn't handle revsets in its subscription config very well and will throw an error if e.g. a branch doesn't exist. - Cloning static HTTP repos may try to stat() a non-existent obsstore file. The issue is that we now care about obsolescence during clone, but statichttpvfs doesn't implement a stat method, so a regular vfs.stat() is used, and it assumes that file is local and calls os.stat(). During a clone, we're trying to stat() .hg/store/obsstore, but in static HTTP case we provide a literal URL to the obsstore file on the remote as if it were a local file path. On windows it actually results in a failure in test-static-http.t. The first issue is going to be addressed in a series dedicated to making sure branchcache is properly and timely written on disk (it wasn't perfect even before this patch, but there aren't enough tests to demonstrate that). The second issue will be addressed in a future patch for notify extension that will make it not raise an exception if a branch doesn't exist. And the third one was partially addressed in the previous patch in this series and will be properly fixed in a future patch when this series is accepted. filteredhash() grows a keyword argument to make sure that branchcache is also invalidated when there are new obsolete revisions in its repo view. This way the on-disk cache format is unchanged and compatible between versions (although it will obviously be recomputed when switching versions before/after this patch and the repo has obsolete revisions). There's one test that uses plain `hg up` without arguments while updated to a pruned commit. To make this test pass, simply return current working directory parent. Later in this series this code will be replaced by what prune command does: updating to the closest non-obsolete ancestor. Test changes: test-branch-change.t: update branch head and cache update message. The head of default listed in hg heads is changed because revision 2 was rewritten as 7, and 1 is the closest ancestor on the same branch, so it's the head of default now. The cache invalidation message appears now because of the cache hash change, since we're now accounting for obsolete revisions. Here's some context: "served.hidden" repo filter means everything is visible (no filtered revisions), so before this series branch2-served.hidden file would not contain any cache hash, only revnum and node. Now it also has a hash when there are obsolete changesets in the repo. The command that the message appears for is changing branch of 5 and 6, which are now obsolete, so the cache hash changes. In general, when cache is simply out-of-date, it can be updated using the old version as a base. But if cache hash differs, then the cache for that particular repo filter is recomputed (at least with the current implementation). This is what happens here. test-obsmarker-template.t: the pull reports 2 heads changed, but after that the repo correctly sees only 1. The new message could be better, but it's still an improvement over the previous one where hg pull suggested merging with an obsolete revision. test-obsolete.t: we can see these revisions in hg log --hidden, but they shouldn't be considered heads even with --hidden. test-rebase-obsolete{,2}.t: there were new heads created previously after making new orphan changesets, but they weren't detected. Now we are properly detecting and reporting them. test-rebase-obsolete4.t: there's only one head now because the other head is pruned and was falsely reported before. test-static-http.t: add obsstore to the list of requested files. This file doesn't exist on the remotes, but clients want it anyway (they get 404). This is fine, because there are other nonexistent files that clients request, like .hg/bookmarks or .hg/cache/tags2-served. Differential Revision: https://phab.mercurial-scm.org/D12097

# verify.py - repository integrity checking for Mercurial
#
# Copyright 2006, 2007 Olivia Mackall <olivia@selenic.com>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.

from __future__ import absolute_import

import os

from .i18n import _
from .node import short
from .utils import stringutil

from . import (
    error,
    pycompat,
    revlog,
    util,
)

VERIFY_DEFAULT = 0
VERIFY_FULL = 1


def verify(repo, level=None):
    with repo.lock():
        v = verifier(repo, level)
        return v.verify()


def _normpath(f):
    # under hg < 2.4, convert didn't sanitize paths properly, so a
    # converted repo may contain repeated slashes
    while b'//' in f:
        f = f.replace(b'//', b'/')
    return f


HINT_FNCACHE = _(
    b'hint: run "hg debugrebuildfncache" to recover from corrupt fncache\n'
)

WARN_PARENT_DIR_UNKNOWN_REV = _(
    b"parent-directory manifest refers to unknown revision %s"
)

WARN_UNKNOWN_COPY_SOURCE = _(
    b"warning: copy source of '%s' not in parents of %s"
)

WARN_NULLID_COPY_SOURCE = _(
    b"warning: %s@%s: copy source revision is nullid %s:%s\n"
)


class verifier(object):
    def __init__(self, repo, level=None):
        self.repo = repo.unfiltered()
        self.ui = repo.ui
        self.match = repo.narrowmatch()
        if level is None:
            level = VERIFY_DEFAULT
        self._level = level
        self.badrevs = set()
        self.errors = 0
        self.warnings = 0
        self.havecl = len(repo.changelog) > 0
        self.havemf = len(repo.manifestlog.getstorage(b'')) > 0
        self.revlogv1 = repo.changelog._format_version != revlog.REVLOGV0
        self.lrugetctx = util.lrucachefunc(repo.unfiltered().__getitem__)
        self.refersmf = False
        self.fncachewarned = False
        # developer config: verify.skipflags
        self.skipflags = repo.ui.configint(b'verify', b'skipflags')
        self.warnorphanstorefiles = True

    def _warn(self, msg):
        """record a "warning" level issue"""
        self.ui.warn(msg + b"\n")
        self.warnings += 1

    def _err(self, linkrev, msg, filename=None):
        """record a "error" level issue"""
        if linkrev is not None:
            self.badrevs.add(linkrev)
            linkrev = b"%d" % linkrev
        else:
            linkrev = b'?'
        msg = b"%s: %s" % (linkrev, msg)
        if filename:
            msg = b"%s@%s" % (filename, msg)
        self.ui.warn(b" " + msg + b"\n")
        self.errors += 1

    def _exc(self, linkrev, msg, inst, filename=None):
        """record exception raised during the verify process"""
        fmsg = stringutil.forcebytestr(inst)
        if not fmsg:
            fmsg = pycompat.byterepr(inst)
        self._err(linkrev, b"%s: %s" % (msg, fmsg), filename)

    def _checkrevlog(self, obj, name, linkrev):
        """verify high level property of a revlog

        - revlog is present,
        - revlog is non-empty,
        - sizes (index and data) are correct,
        - revlog's format version is correct.
        """
        if not len(obj) and (self.havecl or self.havemf):
            self._err(linkrev, _(b"empty or missing %s") % name)
            return

        d = obj.checksize()
        if d[0]:
            self._err(None, _(b"data length off by %d bytes") % d[0], name)
        if d[1]:
            self._err(None, _(b"index contains %d extra bytes") % d[1], name)

        if obj._format_version != revlog.REVLOGV0:
            if not self.revlogv1:
                self._warn(_(b"warning: `%s' uses revlog format 1") % name)
        elif self.revlogv1:
            self._warn(_(b"warning: `%s' uses revlog format 0") % name)

    def _checkentry(self, obj, i, node, seen, linkrevs, f):
        """verify a single revlog entry

        arguments are:
        - obj:      the source revlog
        - i:        the revision number
        - node:     the revision node id
        - seen:     nodes previously seen for this revlog
        - linkrevs: [changelog-revisions] introducing "node"
        - f:        string label ("changelog", "manifest", or filename)

        Performs the following checks:
        - linkrev points to an existing changelog revision,
        - linkrev points to a changelog revision that introduces this revision,
        - linkrev points to the lowest of these changesets,
        - both parents exist in the revlog,
        - the revision is not duplicated.

        Return the linkrev of the revision (or None for changelog's revisions).
        """
        lr = obj.linkrev(obj.rev(node))
        if lr < 0 or (self.havecl and lr not in linkrevs):
            if lr < 0 or lr >= len(self.repo.changelog):
                msg = _(b"rev %d points to nonexistent changeset %d")
            else:
                msg = _(b"rev %d points to unexpected changeset %d")
            self._err(None, msg % (i, lr), f)
            if linkrevs:
                if f and len(linkrevs) > 1:
                    try:
                        # attempt to filter down to real linkrevs
                        linkrevs = []
                        for lr in linkrevs:
                            if self.lrugetctx(lr)[f].filenode() == node:
                                linkrevs.append(lr)
                    except Exception:
                        pass
                msg = _(b" (expected %s)")
                msg %= b" ".join(map(pycompat.bytestr, linkrevs))
                self._warn(msg)
            lr = None  # can't be trusted

        try:
            p1, p2 = obj.parents(node)
            if p1 not in seen and p1 != self.repo.nullid:
                msg = _(b"unknown parent 1 %s of %s") % (short(p1), short(node))
                self._err(lr, msg, f)
            if p2 not in seen and p2 != self.repo.nullid:
                msg = _(b"unknown parent 2 %s of %s") % (short(p2), short(node))
                self._err(lr, msg, f)
        except Exception as inst:
            self._exc(lr, _(b"checking parents of %s") % short(node), inst, f)

        if node in seen:
            self._err(lr, _(b"duplicate revision %d (%d)") % (i, seen[node]), f)
        seen[node] = i
        return lr

    def verify(self):
        """verify the content of the Mercurial repository

        This method run all verifications, displaying issues as they are found.

        return 1 if any error have been encountered, 0 otherwise."""
        # initial validation and generic report
        repo = self.repo
        ui = repo.ui
        if not repo.url().startswith(b'file:'):
            raise error.Abort(_(b"cannot verify bundle or remote repos"))

        if os.path.exists(repo.sjoin(b"journal")):
            ui.warn(_(b"abandoned transaction found - run hg recover\n"))

        if ui.verbose or not self.revlogv1:
            ui.status(
                _(b"repository uses revlog format %d\n")
                % (self.revlogv1 and 1 or 0)
            )

        # data verification
        mflinkrevs, filelinkrevs = self._verifychangelog()
        filenodes = self._verifymanifest(mflinkrevs)
        del mflinkrevs
        self._crosscheckfiles(filelinkrevs, filenodes)
        totalfiles, filerevisions = self._verifyfiles(filenodes, filelinkrevs)

        # final report
        ui.status(
            _(b"checked %d changesets with %d changes to %d files\n")
            % (len(repo.changelog), filerevisions, totalfiles)
        )
        if self.warnings:
            ui.warn(_(b"%d warnings encountered!\n") % self.warnings)
        if self.fncachewarned:
            ui.warn(HINT_FNCACHE)
        if self.errors:
            ui.warn(_(b"%d integrity errors encountered!\n") % self.errors)
            if self.badrevs:
                msg = _(b"(first damaged changeset appears to be %d)\n")
                msg %= min(self.badrevs)
                ui.warn(msg)
            return 1
        return 0

    def _verifychangelog(self):
        """verify the changelog of a repository

        The following checks are performed:
        - all of `_checkrevlog` checks,
        - all of `_checkentry` checks (for each revisions),
        - each revision can be read.

        The function returns some of the data observed in the changesets as a
        (mflinkrevs, filelinkrevs) tuples:
        - mflinkrevs:   is a { manifest-node -> [changelog-rev] } mapping
        - filelinkrevs: is a { file-path -> [changelog-rev] } mapping

        If a matcher was specified, filelinkrevs will only contains matched
        files.
        """
        ui = self.ui
        repo = self.repo
        match = self.match
        cl = repo.changelog

        ui.status(_(b"checking changesets\n"))
        mflinkrevs = {}
        filelinkrevs = {}
        seen = {}
        self._checkrevlog(cl, b"changelog", 0)
        progress = ui.makeprogress(
            _(b'checking'), unit=_(b'changesets'), total=len(repo)
        )
        for i in repo:
            progress.update(i)
            n = cl.node(i)
            self._checkentry(cl, i, n, seen, [i], b"changelog")

            try:
                changes = cl.read(n)
                if changes[0] != self.repo.nullid:
                    mflinkrevs.setdefault(changes[0], []).append(i)
                    self.refersmf = True
                for f in changes[3]:
                    if match(f):
                        filelinkrevs.setdefault(_normpath(f), []).append(i)
            except Exception as inst:
                self.refersmf = True
                self._exc(i, _(b"unpacking changeset %s") % short(n), inst)
        progress.complete()
        return mflinkrevs, filelinkrevs

    def _verifymanifest(
        self, mflinkrevs, dir=b"", storefiles=None, subdirprogress=None
    ):
        """verify the manifestlog content

        Inputs:
        - mflinkrevs:     a {manifest-node -> [changelog-revisions]} mapping
        - dir:            a subdirectory to check (for tree manifest repo)
        - storefiles:     set of currently "orphan" files.
        - subdirprogress: a progress object

        This function checks:
        * all of `_checkrevlog` checks (for all manifest related revlogs)
        * all of `_checkentry` checks (for all manifest related revisions)
        * nodes for subdirectory exists in the sub-directory manifest
        * each manifest entries have a file path
        * each manifest node refered in mflinkrevs exist in the manifest log

        If tree manifest is in use and a matchers is specified, only the
        sub-directories matching it will be verified.

        return a two level mapping:
            {"path" -> { filenode -> changelog-revision}}

        This mapping primarily contains entries for every files in the
        repository. In addition, when tree-manifest is used, it also contains
        sub-directory entries.

        If a matcher is provided, only matching paths will be included.
        """
        repo = self.repo
        ui = self.ui
        match = self.match
        mfl = self.repo.manifestlog
        mf = mfl.getstorage(dir)

        if not dir:
            self.ui.status(_(b"checking manifests\n"))

        filenodes = {}
        subdirnodes = {}
        seen = {}
        label = b"manifest"
        if dir:
            label = dir
            revlogfiles = mf.files()
            storefiles.difference_update(revlogfiles)
            if subdirprogress:  # should be true since we're in a subdirectory
                subdirprogress.increment()
        if self.refersmf:
            # Do not check manifest if there are only changelog entries with
            # null manifests.
            self._checkrevlog(mf._revlog, label, 0)
        progress = ui.makeprogress(
            _(b'checking'), unit=_(b'manifests'), total=len(mf)
        )
        for i in mf:
            if not dir:
                progress.update(i)
            n = mf.node(i)
            lr = self._checkentry(mf, i, n, seen, mflinkrevs.get(n, []), label)
            if n in mflinkrevs:
                del mflinkrevs[n]
            elif dir:
                msg = _(b"%s not in parent-directory manifest") % short(n)
                self._err(lr, msg, label)
            else:
                self._err(lr, _(b"%s not in changesets") % short(n), label)

            try:
                mfdelta = mfl.get(dir, n).readdelta(shallow=True)
                for f, fn, fl in mfdelta.iterentries():
                    if not f:
                        self._err(lr, _(b"entry without name in manifest"))
                    elif f == b"/dev/null":  # ignore this in very old repos
                        continue
                    fullpath = dir + _normpath(f)
                    if fl == b't':
                        if not match.visitdir(fullpath):
                            continue
                        sdn = subdirnodes.setdefault(fullpath + b'/', {})
                        sdn.setdefault(fn, []).append(lr)
                    else:
                        if not match(fullpath):
                            continue
                        filenodes.setdefault(fullpath, {}).setdefault(fn, lr)
            except Exception as inst:
                self._exc(lr, _(b"reading delta %s") % short(n), inst, label)
            if self._level >= VERIFY_FULL:
                try:
                    # Various issues can affect manifest. So we read each full
                    # text from storage. This triggers the checks from the core
                    # code (eg: hash verification, filename are ordered, etc.)
                    mfdelta = mfl.get(dir, n).read()
                except Exception as inst:
                    msg = _(b"reading full manifest %s") % short(n)
                    self._exc(lr, msg, inst, label)

        if not dir:
            progress.complete()

        if self.havemf:
            # since we delete entry in `mflinkrevs` during iteration, any
            # remaining entries are "missing". We need to issue errors for them.
            changesetpairs = [(c, m) for m in mflinkrevs for c in mflinkrevs[m]]
            for c, m in sorted(changesetpairs):
                if dir:
                    self._err(c, WARN_PARENT_DIR_UNKNOWN_REV % short(m), label)
                else:
                    msg = _(b"changeset refers to unknown revision %s")
                    msg %= short(m)
                    self._err(c, msg, label)

        if not dir and subdirnodes:
            self.ui.status(_(b"checking directory manifests\n"))
            storefiles = set()
            subdirs = set()
            revlogv1 = self.revlogv1
            undecodable = []
            for t, f, size in repo.store.datafiles(undecodable=undecodable):
                if (size > 0 or not revlogv1) and f.startswith(b'meta/'):
                    storefiles.add(_normpath(f))
                    subdirs.add(os.path.dirname(f))
            for f in undecodable:
                self._err(None, _(b"cannot decode filename '%s'") % f)
            subdirprogress = ui.makeprogress(
                _(b'checking'), unit=_(b'manifests'), total=len(subdirs)
            )

        for subdir, linkrevs in pycompat.iteritems(subdirnodes):
            subdirfilenodes = self._verifymanifest(
                linkrevs, subdir, storefiles, subdirprogress
            )
            for f, onefilenodes in pycompat.iteritems(subdirfilenodes):
                filenodes.setdefault(f, {}).update(onefilenodes)

        if not dir and subdirnodes:
            assert subdirprogress is not None  # help pytype
            subdirprogress.complete()
            if self.warnorphanstorefiles:
                for f in sorted(storefiles):
                    self._warn(_(b"warning: orphan data file '%s'") % f)

        return filenodes

    def _crosscheckfiles(self, filelinkrevs, filenodes):
        repo = self.repo
        ui = self.ui
        ui.status(_(b"crosschecking files in changesets and manifests\n"))

        total = len(filelinkrevs) + len(filenodes)
        progress = ui.makeprogress(
            _(b'crosschecking'), unit=_(b'files'), total=total
        )
        if self.havemf:
            for f in sorted(filelinkrevs):
                progress.increment()
                if f not in filenodes:
                    lr = filelinkrevs[f][0]
                    self._err(lr, _(b"in changeset but not in manifest"), f)

        if self.havecl:
            for f in sorted(filenodes):
                progress.increment()
                if f not in filelinkrevs:
                    try:
                        fl = repo.file(f)
                        lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]])
                    except Exception:
                        lr = None
                    self._err(lr, _(b"in manifest but not in changeset"), f)

        progress.complete()

    def _verifyfiles(self, filenodes, filelinkrevs):
        repo = self.repo
        ui = self.ui
        lrugetctx = self.lrugetctx
        revlogv1 = self.revlogv1
        havemf = self.havemf
        ui.status(_(b"checking files\n"))

        storefiles = set()
        undecodable = []
        for t, f, size in repo.store.datafiles(undecodable=undecodable):
            if (size > 0 or not revlogv1) and f.startswith(b'data/'):
                storefiles.add(_normpath(f))
        for f in undecodable:
            self._err(None, _(b"cannot decode filename '%s'") % f)

        state = {
            # TODO this assumes revlog storage for changelog.
            b'expectedversion': self.repo.changelog._format_version,
            b'skipflags': self.skipflags,
            # experimental config: censor.policy
            b'erroroncensored': ui.config(b'censor', b'policy') == b'abort',
        }

        files = sorted(set(filenodes) | set(filelinkrevs))
        revisions = 0
        progress = ui.makeprogress(
            _(b'checking'), unit=_(b'files'), total=len(files)
        )
        for i, f in enumerate(files):
            progress.update(i, item=f)
            try:
                linkrevs = filelinkrevs[f]
            except KeyError:
                # in manifest but not in changelog
                linkrevs = []

            if linkrevs:
                lr = linkrevs[0]
            else:
                lr = None

            try:
                fl = repo.file(f)
            except error.StorageError as e:
                self._err(lr, _(b"broken revlog! (%s)") % e, f)
                continue

            for ff in fl.files():
                try:
                    storefiles.remove(ff)
                except KeyError:
                    if self.warnorphanstorefiles:
                        msg = _(b" warning: revlog '%s' not in fncache!")
                        self._warn(msg % ff)
                        self.fncachewarned = True

            if not len(fl) and (self.havecl or self.havemf):
                self._err(lr, _(b"empty or missing %s") % f)
            else:
                # Guard against implementations not setting this.
                state[b'skipread'] = set()
                state[b'safe_renamed'] = set()

                for problem in fl.verifyintegrity(state):
                    if problem.node is not None:
                        linkrev = fl.linkrev(fl.rev(problem.node))
                    else:
                        linkrev = None

                    if problem.warning:
                        self._warn(problem.warning)
                    elif problem.error:
                        linkrev_msg = linkrev if linkrev is not None else lr
                        self._err(linkrev_msg, problem.error, f)
                    else:
                        raise error.ProgrammingError(
                            b'problem instance does not set warning or error '
                            b'attribute: %s' % problem.msg
                        )

            seen = {}
            for i in fl:
                revisions += 1
                n = fl.node(i)
                lr = self._checkentry(fl, i, n, seen, linkrevs, f)
                if f in filenodes:
                    if havemf and n not in filenodes[f]:
                        self._err(lr, _(b"%s not in manifests") % (short(n)), f)
                    else:
                        del filenodes[f][n]

                if n in state[b'skipread'] and n not in state[b'safe_renamed']:
                    continue

                # check renames
                try:
                    # This requires resolving fulltext (at least on revlogs,
                    # though not with LFS revisions). We may want
                    # ``verifyintegrity()`` to pass a set of nodes with
                    # rename metadata as an optimization.
                    rp = fl.renamed(n)
                    if rp:
                        if lr is not None and ui.verbose:
                            ctx = lrugetctx(lr)
                            if not any(rp[0] in pctx for pctx in ctx.parents()):
                                self._warn(WARN_UNKNOWN_COPY_SOURCE % (f, ctx))
                        fl2 = repo.file(rp[0])
                        if not len(fl2):
                            m = _(b"empty or missing copy source revlog %s:%s")
                            self._err(lr, m % (rp[0], short(rp[1])), f)
                        elif rp[1] == self.repo.nullid:
                            msg = WARN_NULLID_COPY_SOURCE
                            msg %= (f, lr, rp[0], short(rp[1]))
                            ui.note(msg)
                        else:
                            fl2.rev(rp[1])
                except Exception as inst:
                    self._exc(
                        lr, _(b"checking rename of %s") % short(n), inst, f
                    )

            # cross-check
            if f in filenodes:
                fns = [(v, k) for k, v in pycompat.iteritems(filenodes[f])]
                for lr, node in sorted(fns):
                    msg = _(b"manifest refers to unknown revision %s")
                    self._err(lr, msg % short(node), f)
        progress.complete()

        if self.warnorphanstorefiles:
            for f in sorted(storefiles):
                self._warn(_(b"warning: orphan data file '%s'") % f)

        return len(files), revisions